Since some years ago in Abruzzo (Central Italy), through a number of monitoring campaigns, a set of more than 1900 indoor radon measures has been acquired by the Regional Agency for the environmental protection. Thus, on the basis of these public experimental data, different statistical approaches, aimed to estimate the probability to exceed the level of 200 Bq/m3 (lower than 300 Bq/m3, threshold value currently recommended by the Euratom commission for indoor radon risk acceptability), taken just as a working reference value, have been selected and discussed in this paper. Essentially, 'Monte Carlo Empirical Bayesian approach', 'Bootstrap' and 'Gibbs samplers' methods have been applied and the results have been partially compared. Moreover, some insights on the minimum number of samples, needed to assess the probability distribution as reasonable as possible, are provided.
Probabilistic estimate of indoor radon distribution in Abruzzo (central Italy): Comparison of different statistical methods / Pasculli, A.; Rizzo, F.; Zazzini, P.. - In: JOURNAL OF PHYSICS. CONFERENCE SERIES. - ISSN 1742-6588. - 1391:1(2019), p. 012001.012001. (Intervento presentato al convegno 8th International Conference on Mathematical Modeling in Physical Science, IC-MSQUARE 2019 tenutosi a svk nel 2019) [10.1088/1742-6596/1391/1/012001].
Probabilistic estimate of indoor radon distribution in Abruzzo (central Italy): Comparison of different statistical methods
Rizzo F.
;
2019-01-01
Abstract
Since some years ago in Abruzzo (Central Italy), through a number of monitoring campaigns, a set of more than 1900 indoor radon measures has been acquired by the Regional Agency for the environmental protection. Thus, on the basis of these public experimental data, different statistical approaches, aimed to estimate the probability to exceed the level of 200 Bq/m3 (lower than 300 Bq/m3, threshold value currently recommended by the Euratom commission for indoor radon risk acceptability), taken just as a working reference value, have been selected and discussed in this paper. Essentially, 'Monte Carlo Empirical Bayesian approach', 'Bootstrap' and 'Gibbs samplers' methods have been applied and the results have been partially compared. Moreover, some insights on the minimum number of samples, needed to assess the probability distribution as reasonable as possible, are provided.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
